A major crisis of the 1980s that shook the world’s financial markets was the 1987 U.S.
market crash. Rigobon (2002a) refers to the major financial crises of the 1990s. The Tequila Effect of the Mexican peso of December 1994, the Asian Flu or “yellow fever” at the end of 1997, the Russian Cold of August 1998, the Brazilian Sneeze of January 1999, and the Nasdaq Rash that began in April 2000. Results on the existence of contagion are mixed. In summary, empirical research based on correlation coefficients produce the same result: During periods
2 During flight-to-quality episodes the risk premium investors require per unit of volatility increases, as
their risk aversion increases (see Vayanos, 2004 for an interesting discussion).
8
of turmoil, correlations tend to go up significantly, pointing toward the existence of a contagion effect.
Various empirical studies show that correlations increase in stock markets during hectic periods and contagious effects occur (Wälti, 2003; Corsetti et al., 2005; Billio and Caporin, 2010; Hossein and Nossman, 2011). Cross-asset correlations generally decrease in times of crises, especially in the case between bonds and stocks (Hunter and Simon, 2004; Connolly et al., 2005). This result can be explained by the flight-to-quality episodes that take place leading to “decoupling”, where high positive correlations among stock markets are observed, but negative correlations between stock and bond markets (Gulko, 2002).
King and Wadhwani (1990) and Lee and Kim (1993) find a significant increase in the cross-country correlation coefficients of returns during the October 1987 U.S. market crash, which is translated as evidence of contagion. The studies by Hamao et al. (1990) and Lin et al.
(1994) also find their roots in the stock market crash of October 1987. Both studies investigate the extent of price volatility and the correlation degree between volatility and returns in New York, Tokyo, and London, and find evidence of contagion across equity markets.
The investigation of mean and volatility spillovers across international developed and emerging stock markets has provided useful insights. Studies such as those by Koutmos and Booth (1995), Ng (2000), and Worthington and Higgs (2004), suggest that spillovers mainly move from developed to emerging markets, and that emerging markets are more integrated than the developed ones. Masson (1998) refers to an effect known as monsoonal where countries are affected simultaneously by crises caused by common shocks, which in turn causes a withdrawal of offshore funds. This simultaneous movement among countries and markets can be explained by common external factors, such as a rise in U.S. interest rates or a devaluation of the dollar, as well as trade linkages and market sentiments.
Calvo and Reinhart (1996) provide evidence of contagion during the Mexican crisis, as they find increased correlations across stock and bond returns for emerging markets in Latin America.
Baig and Goldfajn (1999) suggest discernible patterns of contagion during the East Asian crises and present evidence in favor of substantial contagion in the foreign debt markets, as well as a more tentative evidence on stock market contagion. Corsetti et al. (2005) find evidence of contagion for a small number of countries during the East Asian crisis. Cerra and Saxena (2002) investigate the reasons behind the currency crisis in Indonesia in 1997 and provide evidence that the crisis was a result of contagion from speculative pressures in Thailand and Korea. Glick and Rose (1999) identified currency market contagion across five Asian countries and show that the primary channel of contagion was the strong trade linkages among countries. A similar result is provided by Van Rijckeghem and Weder (2001).
Dungey et al. (2002) examine the transmission of the Russian crisis and the LTCM near-collapse to 12 countries among several world regions, employing the daily behavior of the risk premia in those countries. The results show that there exists significant contagion from both crisis events to other economies in the sample. The LTCM near-collapse appears to have had a larger effect than the Russian crisis on most of the countries. The unscaled by the level of volatility results on whether contagion is more substantial for developed or emerging markets are mixed. Emerging markets such as Brazil and Thailand were more affected by contagion than the U.K., however, Indonesia, Mexico and Korea were less
9
affected by contagion than the U.S. and the Netherlands. On the other hand, the scaled by the level of volatility results show that the magnitude of contagion is relatively larger for emerging countries.
Gravelle et al. (2003) focus on shift contagion and develop a methodology to detect it statistically. In particular, they examine whether existing linkages between assets of different countries remain stable during crises, or whether they grow stronger. They conduct their analysis on the bond markets of four emerging countries and on the currency markets of seven developed countries and provide evidence of shift contagion among these assets.
Their empirical results suggest that, for Latin-American countries, shocks are transmitted via long-term linkages between countries, so that longer-term strategies to deal with contagion might be more effective. Also, for developed currency markets, they suggest that shocks are transmitted only during turbulent periods implying that short-term strategies to stabilize markets may be warranted.
Connolly and Wang (2003) investigate the return comovement in international equity markets with a focus on the distinction between economic fundamentals and contagion. In particular, they examine the potential macro news effect based on a data set of macroeconomic news announcements made in the U.S., U.K., and Japan. Their findings suggest that future inquiry on market comovement may focus on the distinction between contagion and trading on private information, rather than public information. Chan-Lau et al.
(2004) introduce global extreme contagion measures constructed from bivariate extremal dependence measures. Their main results suggest that contagion patterns differ within regions and across regions, with Latin America showing a secular increase in contagion, and that only the 1998 Russian and Brazilian crises led to a global increase in contagion.
Dungey and Martin (2004) measure the contribution of contagion to the volatilities of exchange rates during the East Asian currency crisis, using a multifactor model of exchange rates which allows for both time-dependent common and idiosyncratic factors, as well as unanticipated shocks across currency markets. The empirical results show evidence of significant contagion, especially for Indonesia. Dungey and Martin (2007) formulate an empirical model of multiple asset classes across countries, in which spillover and contagion effects are formally specified. The framework is applied to modeling linkages between currency and equity markets during the East Asian financial crisis of 1997–1998. The results show that spillovers have a relatively larger effect on volatility than contagion, but both are statistically significant. Moreover, in a similar study Dungey et al. (2004) show that there is evidence that the transmission of volatility in the East-Asian currency markets to the developed markets in the region is not due to contagion but due to common world factors.
Ito and Hashimoto (2005) find contagion between equity and currency markets. Bohl and Serwa (2005) test whether European stock markets were affected by a range of crises, namely the Asian, Russian, Brazilian, Argentinean, Turkish, and U.S. ones and find no contagion effects among countries and markets, but only interdependence. Caporale et al.
(2005) test for contagion within the East Asian region using a parameter stability test and controlling for three types of bias, resulting from heteroscedasticity, endogeneity and omitted variable. Their findings suggest the existence of contagion within the East Asian region, consistent with crisis-contingent theories of asset market linkages. Baur and Schulze (2005) introduce a new model to analyze financial contagion based on a modified coexceedance measure. They define contagion as the crisis-specific coexceedance not explained by the covariates for different quantiles. Results for daily stock index returns show
10
that some contagion exists and is predictable within and across regions. Also they show that, contagion depends on a regional market return and its volatility and is stronger for extreme negative returns than for extreme positive returns.
Bekaert et al. (2011) analyze the transmission of crises to country-industry equity portfolios in 55 countries using the 2007-2009 financial crisis as a laboratory. They find evidence of systematic contagion from U.S. markets and from the global financial sector although the effects are very small, however, they show that there is substantial contagion from domestic equity markets to individual domestic equity portfolios. Brière et al. (2012) reject contagion for fixed-income assets, detect contagion effects at the 5% level in stocks, and finally conclude that contagion is an artifact caused by globalization. A similar result (but no similar interpretation) has been provided by Forbes and Rigobon (2002), and Rigobon (2002b, 2003) who find little evidence of shift contagion during the Mexican, Asian, and Russian crises in several emerging markets, as well as between 1994 and 1999 in the Argentinean and Mexican bond markets. Instead, they find a continued high level of correlation during calm periods which they interpret as interdependence.
Beirne and Gieck (2012) provide an empirical assessment of interdependence and contagion across bonds, stocks, and currencies for over 60 economies over the period 1998 to 2011.
Their findings indicate that interdependence is most notable across advanced and emerging economies, in the case of the equity market, while contagion effects are most apparent in Latin America and Emerging Asia. However, they also find evidence of contagion from global bonds to regional stocks in Central and Eastern Europe, Middle East and Africa regions.
Interdependence within the bond market applies mainly to the advanced economies, whereas evidence for bond market contagion is found for Mexico, Venezuela and Philippines. Cross-market interdependence and contagion from global equities and global currencies to local bonds is not prevalent. Finally, exchange rate interdependence is important for advanced economies, whereas contagion is present in domestic currencies in Hong Kong, Korea, Thailand, Slovakia and Australia that are susceptible to global currency shocks.
Beirne et al. (2013) apply the concept of shift contagion to the analysis of spillovers from mature to emerging stock markets and test for shifts in the transmission mechanism during episodes of extreme movements in mature markets. Their analysis covers a large sample of 41 emerging market economies in Asia, Europe, Latin America, and the Middle East. They show that spillovers from mature markets influence the dynamics of conditional variances of returns in emerging stock markets, and that the spillover parameters change during turbulent episodes in mature markets.